<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Tekai - Hold Ring</title><description>Catalog entries in the Hold ring of the Tekai technology radar.</description><link>https://tekai.dev/</link><language>en</language><item><title>Agent Communication Protocol (ACP)</title><link>https://tekai.dev/catalog/agent-communication-protocol/</link><guid isPermaLink="true">https://tekai.dev/catalog/agent-communication-protocol/</guid><description>IBM Research&apos;s REST-based open protocol for AI agent interoperability, enabling agents built on different frameworks to discover and communicate with each other via standard HTTP and MIME types; merged into the A2A protocol under the Linux Foundation in August 2025.</description><pubDate>Thu, 23 Apr 2026 00:00:00 GMT</pubDate><category>ai-agents</category><category>interoperability</category><category>protocol</category><category>ibm</category><category>linux-foundation</category><category>rest</category><category>http</category><category>deprecated</category><category>agent-communication</category></item><item><title>Codel</title><link>https://tekai.dev/catalog/codel/</link><guid isPermaLink="true">https://tekai.dev/catalog/codel/</guid><description>Open-source autonomous AI coding agent (2024) that runs inside Docker with a web UI, executing tasks via terminal, browser automation, and a built-in file editor backed by PostgreSQL history.</description><pubDate>Sat, 11 Apr 2026 00:00:00 GMT</pubDate><category>autonomous-agent</category><category>ai-coding-agent</category><category>docker</category><category>sandboxed-execution</category><category>self-hosted</category><category>go</category><category>ollama</category><category>openai</category></item><item><title>Loom</title><link>https://tekai.dev/catalog/loom/</link><guid isPermaLink="true">https://tekai.dev/catalog/loom/</guid><description>A proprietary Rust monorepo by Geoffrey Huntley (creator of the Ralph Loop Pattern) implementing self-hosted infrastructure for LLM-powered agent loops: server-side LLM proxy, Kubernetes-based remote execution (Weaver), full auth stack, and multi-agent observability.</description><pubDate>Sat, 11 Apr 2026 00:00:00 GMT</pubDate><category>rust</category><category>ai-agents</category><category>autonomous-loops</category><category>llm-proxy</category><category>kubernetes</category><category>multi-agent</category><category>agent-infrastructure</category><category>ralph-loop</category><category>proprietary</category></item><item><title>OpenClaw (NemoClaw context)</title><link>https://tekai.dev/catalog/openclaw-nemoclaw-context/</link><guid isPermaLink="true">https://tekai.dev/catalog/openclaw-nemoclaw-context/</guid><description>Note: OpenClaw as a sandboxed agent runtime target for NVIDIA NemoClaw. See frameworks/openclaw.md for the full OpenClaw catalog entry (open-source, MIT, multi-channel gateway).</description><pubDate>Sat, 11 Apr 2026 00:00:00 GMT</pubDate><category>ai-agents</category><category>openclaw</category><category>nvidia</category><category>nemoclaw</category><category>duplicate-note</category></item><item><title>Pi-Builder</title><link>https://tekai.dev/catalog/pi-builder/</link><guid isPermaLink="true">https://tekai.dev/catalog/pi-builder/</guid><description>An early-stage TypeScript monorepo that wraps multiple CLI coding agents and routes tasks between them using a capability declaration system backed by SQLite.</description><pubDate>Sat, 11 Apr 2026 00:00:00 GMT</pubDate><category>ai-coding-agent</category><category>agent-orchestration</category><category>capability-routing</category><category>typescript</category><category>sqlite</category><category>cli</category><category>monorepo</category></item><item><title>smol developer</title><link>https://tekai.dev/catalog/smol-developer/</link><guid isPermaLink="true">https://tekai.dev/catalog/smol-developer/</guid><description>Open-source Python library that generates entire codebases from a natural-language prompt using a three-stage pipeline: shared-dependency manifest, file-path enumeration via function calling, then parallel per-file generation.</description><pubDate>Sat, 11 Apr 2026 00:00:00 GMT</pubDate><category>ai-coding-agent</category><category>codebase-generation</category><category>prompt-to-code</category><category>embeddable-agent</category><category>python</category><category>gpt-4</category><category>scaffolding</category><category>human-in-the-loop</category></item><item><title>MMLU (Massive Multitask Language Understanding)</title><link>https://tekai.dev/catalog/mmlu/</link><guid isPermaLink="true">https://tekai.dev/catalog/mmlu/</guid><description>A benchmark of 15,908 multiple-choice questions across 57 academic subjects for evaluating LLM knowledge, now effectively saturated by frontier models.</description><pubDate>Fri, 03 Apr 2026 00:00:00 GMT</pubDate><category>ai-evaluation</category><category>benchmark</category><category>nlp</category><category>language-models</category><category>saturated</category><category>knowledge-testing</category></item><item><title>Vivaria</title><link>https://tekai.dev/catalog/vivaria/</link><guid isPermaLink="true">https://tekai.dev/catalog/vivaria/</guid><description>METR&apos;s open-source platform for running AI agent evaluations and elicitation research, now deprecated in favor of Inspect AI.</description><pubDate>Fri, 03 Apr 2026 00:00:00 GMT</pubDate><category>ai-safety</category><category>ai-evaluation</category><category>agent-elicitation</category><category>benchmarking</category><category>deprecated</category></item></channel></rss>